Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Would faster time control counter the faster cpu speed

Author: Uri Blass

Date: 15:06:40 02/26/03

Go up one level in this thread


On February 26, 2003 at 17:51:41, John Wentworth2 wrote:

>On February 26, 2003 at 17:42:31, Uri Blass wrote:
>
>>On February 26, 2003 at 17:32:08, Uri Blass wrote:
>>
>>>On February 26, 2003 at 16:56:54, John Wentworth2 wrote:
>>>
>>>>On February 26, 2003 at 16:28:41, James T. Walker wrote:
>>>>
>>>>>Just for fun I tried the Mephisto Risc (1MB) dedicated machine which was a world
>>>>>champion around 1991 against Fritz 8.  The Mephisto was playing at 40/2hours
>>>>>while Fritz 8 played at G/5min.  Some info: The Mephisto searches about 3-4K/nps
>>>>>and is currently rated about 2100 on the SSDF list.  When new it was considered
>>>>>to be rated much higher.  Not sure but I think Ed S. was the programmer.  Here
>>>>>are the 2 games.
>>>>>
>>>>>[Event "Blitz:5'"]
>>>>>[Site "?"]
>>>>>[Date "2003.02.26"]
>>>>>[Round "1"]
>>>>>[White "Mephisto (40/2hrs)"]
>>>>>[Black "Fritz 8"]
>>>>>[Result "0-1"]
>>>>>[ECO "B63"]
>>>>>[PlyCount "92"]
>>>>>
>>>>>1. e4 c5 2. Nf3 Nc6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 d6 6. Bg5 e6 7. Qd2 Be7 8.
>>>>>O-O-O O-O 9. Nb3 a6 10. h4 a5 11. Bxf6 Bxf6 12. Qxd6 Qxd6 13. Rxd6 a4 14. Na1
>>>>>Be5 15. Rd1 a3 16. Rh3 f5 17. exf5 Rxf5 18. bxa3 Rxf2 19. Kb2 Ra4 20. Nb3 Raf4
>>>>>21. Rh1 Ne7 22. Rd8+ Rf8 23. Rxf8+ Rxf8 24. Bc4 b5 25. Bxb5 Nd5 26. Rh3 Nxc3
>>>>>27. Rxc3 Bb7 28. Nc5 Bxg2 29. Kb3 Bd5+ 30. Rc4 Bd6 31. Nd3 Rf5 32. Nb4 Bxc4+
>>>>>33. Bxc4 Kf7 34. Nc6 Bg3 35. Nd8+ Kf6 36. Nxe6 Bxh4 37. Nd4 Re5 38. Bd3 Bf2 39.
>>>>>Nf3 Re7 40. Nd2 Be1 41. Ne4+ Ke5 42. Ng5 h5 43. Nf3+ Kf4 44. Nxe1 Rxe1 45. c4
>>>>>h4 46. Bg6 Resigns 0-1
>>>>>
>>>>>[Event "Blitz:5'"]
>>>>>[Site "?"]
>>>>>[Date "2003.02.26"]
>>>>>[Round "2"]
>>>>>[White "Fritz 8"]
>>>>>[Black "Mephisto (40/2hrs)"]
>>>>>[Result "1-0"]
>>>>>[ECO "B12"]
>>>>>[PlyCount "33"]
>>>>>
>>>>>1. d4 c6 2. e4 d5 3. e5 Bf5 4. Nf3 e6 5. Be2 c5 6. Be3 Qb6 7. Nc3 Qxb2 8. Nb5
>>>>>Na6 9. dxc5 Nb4 10. O-O Nxc2 11. Qd2 Rc8 12. Nxa7 Rc7 13. Qa5 Rd7 14. Bb5 Nxa1
>>>>>15. Bd4 Qxb5 16. Nxb5 Be7 Resigns 1-0
>>>>
>>>>
>>>>I wonder over a period of blitz games if the Mephisto would do better. The
>>>>reason I say this is because I recently played a 2' blitz match between Crafty
>>>>19.01 on a 1.2Ghz PC and Fritz 6 on an old 166Mhz PC. Fritz 6 on the slow
>>>>machine won by a large margin, something like 46.5 - 14.5 but when the time
>>>>control was longer, Crafty played much better.
>>>
>>>46.5-14.5 when the hardware difference is so big.
>>>Something seems to be wrong.
>>>
>>>How did you play the match?
>>>If you used an autoplayer then it is possible that the autoplayer is broken
>>>at this time control.
>>>
>>>Crafty is not a good blitz engine but I believe that it is not going to lose
>>>46.5-14.5 against Fritz6 at any time control if you give it hardware advantage
>>>of more than 5-1.
>>>
>>>Uri
>>
>>I meant to say not a good blitz engine relative to Fritz and the same is for
>>longer time control.
>>
>>I do not know if relatively crafty is weaker in blitz.
>>It may be interesting if people try to find x when
>>
>>Fritz6(1 minutes) has the same strength as Crafty19.01 (x minutes)
>>
>>It is possible to try to find it by doing a match of 800 games from the
>>positions after 1.a3 a6 1.a3 a5,.... and finding x such that the result is the
>>closest to 400-400)
>>
>>After that people can try to find based on similiar match if
>>Fritz6(10 minutes) is weaker or stronger than Crafty(10x minutes).
>>
>>people can use other programs and they can use Fritz8 and I only suggested
>>fritz6 because it was in the discussion.
>>
>>
>>Uri
>
>
>I did use the autoplayer, so maybe it had something to do with that but it
>appeared to play normally. I was quite suprised by the result as well. I was
>surprised that Fritz 6 on the slow machine would win at all but by such a wide
>margin was really shocking to me.

Note that I never use the autoplayer but
I remember that I read that there may be a problem of a small delay of 2 seconds
in getting the moves from the autoplayer(not significant at 120/40
but may be significant at blitz) and the question is if Fritz6 really used 2
minutes per game(for example if fritz6 used more than it when Crafty kept to use
only it's time then there is a problem).

Uri



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.